Progress toward understanding and curing many human neurological diseases is hindered by a lack of methods for objective and quantitative assessment of the connections between genes and behavior. Traditional studies of animal behavior probe only a small subset of user-defined behaviors in low- dimensional data and are inappropriate for the discovery of subtle phenotypes or complex relationships between genes and behavior. To move beyond these simple representations of animal behavior, we propose to build a high-resolution, three-dimensional imaging system that can track single behaving animals. This system will quantitatively connect behavior to genetics using a data-driven method for the discovery of stereotyped motions. Our specific aims are to 1) develop a new standard for behavioral observation and automatic phenotyping to quantify the behavior of freely moving fruit flies, 2) demonstrate the sensitivity and accuracy of our new technology by studying Drosophila models of neurodegenerative diseases and the behavior of closely-related species of Drosophila fruit flies and 3) expand our system to enable simultaneous tracking and analysis of two independent flies in a single arena and test this system with assays of courtship behavior between male and female flies. The ability to measure subtle changes in behavior in a model genetic organism will bring us closer than we have ever been to understanding neurological diseases and behavioral disorders at the genetic and molecular level.